Losing Talent to a Startup Might Be the Best Thing That Ever Happened to Meta

What if the company “losing” the talent war is actually winning something else entirely?

That’s the strange, counterintuitive story playing out right now between Meta and Thinking Machines Lab — a startup that, depending on which week you’re reading this, is either raiding Meta’s talent pool or feeding it. Both things are true at the same time, and that tells us something fascinating about how AI power is shifting in 2025.

Meet Thinking Machines Lab

If you haven’t heard of Thinking Machines Lab yet, here’s your quick intro. The startup was founded by Mira Murati, who previously served as Chief Technology Officer at OpenAI. That pedigree alone was enough to turn heads. But TML has since built a reputation serious enough to attract researchers away from one of the most well-resourced AI labs on the planet — Meta.

The appeal isn’t hard to understand. TML carries a $12 billion valuation, which means early employees are sitting on equity that could be genuinely life-changing. For researchers who want to do ambitious work without the bureaucratic weight of a tech giant, a well-funded startup with a credible founder is a compelling offer.

The Talent Revolving Door

Here’s where the story gets interesting. This isn’t a simple case of a scrappy startup stealing people from a big corporation. The movement goes both ways.

Meta has been pulling talent from Thinking Machines Lab too. Two TML employees — Mark Jen and Yinghai Lu — have made the move over to Meta. So while TML is attracting researchers away from Meta, Meta is simultaneously picking up people from TML. Each company is, in its own way, benefiting from what the other builds.

Think of it less like a war and more like a talent ecosystem. People move, ideas travel with them, and both organizations end up shaped by the exchange. For non-technical readers, a useful analogy: imagine two restaurants on the same street, occasionally trading chefs. Both menus get more interesting over time.

Why the Google Deal Changes Everything

If the talent story is interesting, the infrastructure story is where things get genuinely significant for TML’s future.

Thinking Machines Lab recently signed a multibillion-dollar cloud deal with Google. That’s not just a financial milestone — it’s a statement about what kind of player TML intends to be. As part of that deal, TML gets access to Nvidia’s latest GB300 chips, making it one of the first organizations to work with that hardware.

For anyone wondering why that matters: AI models are only as capable as the hardware they’re trained on. Access to the newest, most powerful chips means TML can train larger, faster, and more capable models than most of its competitors. A startup with top-tier talent and top-tier compute is no longer just a promising name — it’s a serious contender.

What This Means for the Broader AI Space

The Meta-TML dynamic is a useful window into how AI development actually works right now. It’s not a clean hierarchy where big companies sit at the top and startups scramble below them. The lines are blurry, the talent flows in multiple directions, and a well-timed deal or a credible founder can shift the balance quickly.

For everyday people watching this from the outside, a few things are worth keeping in mind:

• Startups with strong founders and solid funding can compete directly with tech giants for the best researchers.
• Talent movement between companies isn’t just about salaries — equity, mission, and the chance to work on something new all play a role.
• Infrastructure deals like TML’s arrangement with Google can close the gap between a startup and an established lab faster than most people expect.

So Who’s Really Winning?

Framing this as a win-lose situation misses the point. Meta loses some researchers to TML, gains others back, and continues operating at massive scale. TML attracts top talent, secures a major cloud deal, and builds momentum. Both organizations are growing stronger through the friction between them.

If anything, the real winner is the broader field of AI research. When talent moves freely and startups can access serious compute, more ideas get tested, more approaches get tried, and the pace of progress picks up across the board.

Mira Murati built something that Meta researchers want to join. Meta built something that TML researchers want to join. That’s not a crisis for either company — that’s just what a healthy, competitive space looks like.

🕒 Published: April 24, 2026